|Publication number||US5230047 A|
|Application number||US 07/510,209|
|Publication date||Jul 20, 1993|
|Filing date||Apr 16, 1990|
|Priority date||Apr 16, 1990|
|Also published as||EP0453193A2, EP0453193A3|
|Publication number||07510209, 510209, US 5230047 A, US 5230047A, US-A-5230047, US5230047 A, US5230047A|
|Inventors||Alexander H. Frey, Jr., Richard C. Mosteller|
|Original Assignee||International Business Machines Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (7), Non-Patent Citations (2), Referenced by (139), Classifications (12), Legal Events (6)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention relates to improvements in parallel computing systems, and, more particularly, to a method for balancing distributed file structures in a manner to enable recovery of data in the event of a failure.
Data structures in computer systems often use tree-organizations to enable search paths to be determined to specific information. A file contains whatever information the user places in it (e.g., an executable program), and may be stored anywhere in the memory of a data processing system. Directories provide the mapping between a file's name or other indicator and the file itself. Within a tree-organized search structure, a "ROOT" is generally found which enables all information to be located by tracing a path through a stated chain of branches, until the desired information is reached. The ROOT of the tree indicates spans of information handled by each of the subservient branches of the next level so that when a specific information search is specified, a search through the ROOT indicates in which subservient branch in the next level further information concerning the search may be found.
Such a tree-structured information array is often used by a key-index data arrangement which is an array of paired entries. The first portion of the entry is a key value and the second is an index value (such as a record number) associated with that key value. A key-index file is accessed by giving the key, with the system returning the key-index pair. This enables files to be accessed by a key name rather than by a record number or byte number.
In a tree-structured information array used by uni-processor computing systems, structures known as balanced trees (i.e., B-trees) have been utilized. B-trees enable search work to be evenly distributed throughout the tree and prevent portions of the tree from being overloaded, while other portions are sparsely populated. A description of B-trees can be found in "The Art of Computer Programming", Volume 3/Sorting and Searching, D. E. Knuth, Addison-Wesley Company, 1973, pp. 473-479.
Tree structured information arrays in parallel computing systems present significantly different problems than those that are present in uni-processor systems. Such parallel computing systems may include a large number of "nodes" that are interconnected by a high speed communication network. Each node generally comprises a processor and a memory, operates independently, and interacts with other nodes via message traffic and transfers of blocks of data over the communication network.
It may be the case in parallel computing systems that file data structures are distributed throughout a plurality of nodes rather than being centrally located in a single node In addition to enabling more efficient diffusion of work among the nodes, such distributed data structures must provide for system recovery in the event of malfunction of one or more nodes. Tree-structured information arrays are one example of files that may be used in such systems and may be distributed among the nodes. When, however, B-tree structures are employed, problems may occur if, during rebalancing nodes of the tree-structure, a node storing a portion of the tree should fail or the entire system should fail. In such case, it is necessary for the system to know (1) whether there was or was not a rebalancing between different nodes in process; (2) what nodes were involved with the rebalancing effort; and (3) at what state was the rebalancing effort when the failure occurred. Knowing these facts, the system, after the failed node has been recovered or the entire system restarted then must reconstruct itself in a manner so as to continue the rebalancing effort, while not losing data in the process.
Parallel computing systems employing tree-oriented data structures may be found in U.S. Pat. Nos. 4,860,201; 4,766,534; 4,412,285; 4,445,171; and U.S. Pat. No. 4,825,354. In the main, those prior art patents are mainly concerned more with operational message transfers between the nodes of a parallel data processing system rather than recovery of data in the event of a failure.
Accordingly, it is an object of this invention to provide a method for data recovery in a parallel computing network.
It is another object of this invention to provide a method for data recovery in a parallel computing network which employs a B-tree information array file structure.
It is a further object of this invention to provide an efficient method for rebalancing a B-tree file data structure across different nodes in a parallel data computing network.
A distributed network is described which contains a plurality of interconnected nodes, each node including a processor and data storage apparatus. A plurality of key-index data structures are distributed among the nodes, with each node including a B-tree data structure defining the locations of key-index data structures on that node. The tree data structure includes a ROOT data structure comprising key-index data identifiers and assigned nodes for the top level of the tree. A method is described for balancing the tree data structure which comprises the steps of:
a. providing in each of the nodes across which the data files are distributed, at least a pair of copies of the ROOT data structure;
b. determining when a first node contains an excess of key-index data identifiers;
c. updating one part (called "NEW ROOT") of the ROOT data structure in the file in non-volatile storage containing the key-index file structure, to note the movement of the key-index data identifiers, whereby in the event of a malfunction of one of the nodes, a record exists in both of the nodes of both an updated and non-updated ROOT data structure to enable data recovery;
d. moving the excess of key-index data identifiers to a second node;
e. updating the second copy (called "OLD ROOT") of the ROOT data structure after all the key-index data identifiers have been moved, to indicate that the rebalancing operation has been completed.
FIG. 1 is a block diagram of a parallel computing system having a three dimensional structure.
FIG. 2 is a block diagram illustrating the structure of each node in the parallel computing system of FIG. 1.
FIG. 3 is a schematic diagram of a B-tree, as configured in a parallel computing system.
FIG. 4 is a schematic diagram of an exemplary ROOT for the B-tree shown in FIG. 3.
FIGS. 5 and 6 illustrate a high level flow diagram showing the B-tree rebalance algorithm of the invention.
FIG. 7 is a logic flow diagram showing system recovery after a power-down.
FIG. 8 is a logic flow diagram illustrating a node's operation when a data insertion or extraction is encountered during rebalancing.
In FIG. 1, the topology of a representative three dimensional parallel computing system is illustrated. Each box 10 represents a node and each line 12 represents a communication link between the nodes. It will be noted that the nodes are connected in the form of a torus and, while multiple levels of the parallel computing system are illustrated, interconnections between levels are omitted from the FIG. From a review of the nodes in level 14, it can be seen that each node is connected to an adjoining node in both the X and Y dimension and furthermore, each node has 2 additional links (not shown), one to the counterpart node in the plane above and one to the counterpart node in the plane below.
As shown in FIG. 2, the structure of each node is similar and comprises a microprocessor 16 which is connected via a communication channel 18 to a buffer memory 20, a device adapter 22 and an I/O adapter and link router 24. Connected to device adapter 22 is a disk memory 26. Each node, as shown in FIG. 2, operates substantially independently of other nodes in processing its portion of a program or subroutine.
Key-index data identifiers employed during the operation of nodes 10 are distributed among all nodes so as to balance the work load therebetween. Furthermore, such a distribution of key-index data identifiers throughout the nodal structure insures that the system can recover in the event of single nodal failure, and in certain cases, multiple nodal failures.
When key-index data identifiers are distributed among a plurality of nodes, a plurality of tree-structured arrays are required to enable location of any particular stored key-index data identifier. An efficient structure for keeping track of identifier locations is a B-tree structure. As indicated in the introduction to this specification, a feature of a B-tree structure is that it may be continually rebalanced to assure that no particular node of the tree becomes more heavily populated with key-index data identifiers than any other node. This assures that no node is overloaded with work as compared to all other nodes.
Algorithms are known in the prior art for constructing B-trees and for maintaining them in a balanced or nearly balanced state in uni-processor systems. Nevertheless, when a B-tree structure is employed in a parallel computing system and is, in itself, distributed among the nodes of the parallel computing system, it is vital that it be structured so that if a B-tree rebalance operation that involves more than one node is in-process, then the system state can be determined in the event of a malfunction and system recovery can occur after the malfunction has been corrected.
In FIG. 3, an illustrative B-tree structure is illustrated. As above described, each entry in the B-tree is a key-index value with the keys generally being listed in sequential order. The uppermost level of the B-tree contains a ROOT directory 31 which indicates, for each span of key values, where in the file system either the next level of directory indications can be found, or where the actual file can be found.
As shown in FIG. 4, the actual structure of a ROOT is two lists, which are initially identical. The first list is a node's "old" initial key value ("OLD" ROOT) and the second list is the node's new initial key value ("NEW" ROOT). A key value of zero is assumed for the first node (Node 1). For instance, in the NEW ROOT list shown in FIG. 4, the first entry is key number 250 and the second is 500. That indicates that key values from 0 to 250 are located in node 1 and that values from 251 to 500 are located in node 2, etc. It is important to realize that only when rebalancing is in process, do the key values in New Root and Old Root differ (i.e. when not rebalancing, old and new initial key values stored in a node are equal).
As shown in FIG. 3, node 2 includes a child branch 30 which in turn, has a plurality of child branches 32. Emanating from each of child branches 32 is a plurality of leafs 34 where the actual key-index file identifiers can be found.
During processing in a parallel computing system, files are mapped into the various nodes in accordance with predetermined characteristics of each file. It can occur, notwithstanding the balanced key-index arrangement which is shown in FIG. 3, that one or more nodes can become unbalanced with respect to another node if too many key-index identifiers are assigned to it. In general, each child branch of the tree is restricted in its fan out, so that as more key-index pairs are assigned to a particular node, additional levels of branching are inserted to accommodate those child branches in which entries overflow the branch's limited fan out.
Within each node, the number of child branch levels can be determined by merely counting the number of key entries. Thus, by one node comparing its number of key entries with the number of key entries in an adjoining node, an unbalance in tree levels can be readily determined and correction instituted. However, if a node is in the process of rebalancing its key entries with an adjacent node and the entire system or the adjacent node malfunctions, the system must be provided with a method and means for data recovery. This recovery method is illustrated in FIGS. 5 and 6.
In FIG. 5, box 50 indicates that initially, each node is provided with certain basic data. As part of that data, each node is provided with a copy of ROOT for the B-tree structure. Each node then duplicates that record and designates one copy as OLD ROOT and another as NEW ROOT. Additionally, each node is provided with a criteria for determining its state of balance or unbalance with respect to an adjacent node. That criteria may be the number of keys present in that node, or a percentage of B-tree branches filled up or any other appropriate criteria.
Neighboring nodes continually exchange information regarding their state with regards to the unbalance/balance criteria. In other words,. messages are frequently passed between adjacent nodes enabling the nodes to detect an unbalanced state with respect to an adjacent node. Each node checks its number of key-index entries and compares that number with the number of entries in an adjacent node. If a node finds that the adjacent node has a lesser number of key-index between the nodes meets a criteria which calls for balancing to commence, the node with "too many" entries commences the transfer action.
For the sake of example, it will be assumed that there are two nodes i.e., node i and node i+1. It will be further assumed that node i detects a tree unbalance (box 52) with respect to adjacent node i+1, in that node i has many more keys than node i+1. (If no unbalance is detected it merely continues exchanging data until it determines that an unbalance does occur). When node i detects an unbalance, it requests to transfer key-indexes to node i+1 (box 54). Node i determines a new ROOT boundary for itself and writes its new boundary value into its "new ROOT". The creation of the new boundary value within "new ROOT"determines how many keys will be moved from node i to node i+1.
For example, in node i the indicated key values in the first child branch may be from 100 to 145 whereas in node i+1 the key values may only span from 146 to 151. As a result, node i requests to transfer key values 126 to 146 to node i+1, thereby causing a readjustment in the node i lower boundary from 146 to 126. To do this, Node i writes 126 into its list of boundary values called "NEW ROOT". Thus, by establishing, the new boundary value of 126, node i indicates that it will be transferring keys between 126 and 146 to node i+1. At this point, node i sets an internal pointer equal to its current boundary value, say, for instance, key-value 145 (box 58).
Then, node i moves the key at the current boundary value (145) from itself to node i+1 (box 60). Node i+1, upon receiving the key and storing it to disk, acknowledges receipt of the key (decision box 62) and updates its copy of New Root to indicate the newly received key.
If node i does not receive a move acknowledgement from node i+1, it repeats the attempt to move the key (decision box 64). If n retries have not occurred, the program recycles to box 60 until n tries have occurred, at which point node i issues an abort indication.
Assuming that the key move is acknowledged, node i deletes that key from its tree (box 66), and moves the pointer to the next key value, e.g., 144, (box 68). Node i then determines whether the that key value is less than its NEW ROOT boundary (decision box 70). If the answer is no, the key is transferred, and the algorithm recycles back to box 60. If the value of the key is found to be less than the NEW ROOT, the transfer process is complete and the algorithm proceeds to box 72 where both nodes i and i+1 write the new ROOT value over old ROOT. At this stage, a broadcast is made to other nodes of the value of new ROOT so that other nodes can then update their local copes of the values of ROOT.
It can be seen from the above, that when node i starts sending keys, one by one, to node i+1, node i requests node i+1 to insert those keys in its part of the B-tree, but node i does not immediately remove those keys from its part of the B-tree. When insertions occur into the B-tree in node i+1, the temporary boundary pointer in node i+1 and the temporary boundary pointer in node i change to reflect those insertions. During this period, other nodes can still access the part of the B-tree belonging to nodes i and nodes i+1 since nothing has been deleted from either node. At the same time, new ROOT reflects insertions are being made into node i+1. If either of the nodes goes down during this period, it can recover when it comes up and sees that new ROOT and old ROOT have different values For instance, see FIG. 7 wherein at power up, each node compares its OLD ROOT values to its NEW ROOT values (box 100). If the values are the same (decision box 102) then the node continues and no rebalancing is required. On the other hand, if the values are determined to be different, the node selects its key closest to the OLD ROOT boundary (box 104) and continues the rebalancing process as shown at boxes 60 et sequence in FIGS. 5 and 6.
As above stated, once all of the keys that are being moved become inserted in the B-tree of node i+1, node i+1 overwrites its OLD ROOT value with the NEW ROOT value and propagates its NEW ROOT value throughout the system. Once the NEW ROOT value is propagated to other nodes, all queries related to the part of the tree that has been moved are addressed to node i+1. However, if a key insertion or deletion from another node occurs during rebalancing, the procedure shown in FIG. 8 occurs. The value of the key being inserted is compared with the pointer to the next key to be transferred. If the key value is found to be larger (box 112), then the node transfers the key to the node having the proper span of key values. If, on the other hand, it is found to be smaller in value than the next key to be transferred, then the node inserts the key in its B-tree (box 114).
To summarize, the pointer separating the part of the B-tree that has already been copied to another node from the part that still has not been copied enables any insertion or deletion to be properly directed. Thus, whenever an insertion or deletion operation is received at a node, it is forwarded to a new node if the part of the B-tree that belongs to the node receiving the insertion or deletion instruction has been copied to the new node.
As can now be seen, B-tree balancing results in B-trees being in a continuous rebalancing state, i.e., they exchange information regarding their unbalanced criteria on a continuous basis and whenever one node has been overfilled, they initiate rebalancing. Furthermore, if any node crashes during the rebalancing procedure, the values of NEW ROOT and OLD ROOT enable it to either continue its rebalancing effort or not, as the case may be.
It should be understood that the foregoing description is only illustrative of the invention. Various alternatives and modifications can be devised by those skilled in the art without departing from the invention. Accordingly, the present invention is intended to embrace all such alternatives, modifications and variances which fall within the scope of the appended claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4412285 *||Apr 1, 1981||Oct 25, 1983||Teradata Corporation||Multiprocessor intercommunication system and method|
|US4445171 *||Apr 1, 1981||Apr 24, 1984||Teradata Corporation||Data processing systems and methods|
|US4575798 *||Jun 3, 1983||Mar 11, 1986||International Business Machines Corporation||External sorting using key value distribution and range formation|
|US4766534 *||Oct 16, 1986||Aug 23, 1988||American Telephone And Telegraph Company, At&T Bell Laboratories||Parallel processing network and method|
|US4825354 *||Nov 12, 1985||Apr 25, 1989||American Telephone And Telegraph Company, At&T Bell Laboratories||Method of file access in a distributed processing computer network|
|US4860201 *||Sep 2, 1986||Aug 22, 1989||The Trustees Of Columbia University In The City Of New York||Binary tree parallel processor|
|US5020059 *||Mar 31, 1989||May 28, 1991||At&T Bell Laboratories||Reconfigurable signal processor|
|1||Knuth, "The Art of Computer Programming", vol. 3 Sorting and Searching, Addison-Wesley Publishing Co., 1973, pp. 473-479.|
|2||*||Knuth, The Art of Computer Programming , vol. 3 Sorting and Searching, Addison Wesley Publishing Co., 1973, pp. 473 479.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5418945 *||May 18, 1992||May 23, 1995||Motorola, Inc.||File based and highly available hybrid database|
|US5546539 *||Oct 12, 1995||Aug 13, 1996||Intel Corporation||Method and system for updating files of a plurality of storage devices through propogation of files over a nework|
|US5634107 *||Oct 6, 1995||May 27, 1997||Fujitsu Limited||Data processor and method of processing data in parallel|
|US5675736 *||Jul 24, 1996||Oct 7, 1997||International Business Machines Corporation||Multi-node network with internode switching performed within processor nodes, each node separately processing data and control messages|
|US5805158 *||Aug 22, 1996||Sep 8, 1998||International Business Machines Corporation||Copying predicted input between computer systems|
|US5806065 *||May 6, 1996||Sep 8, 1998||Microsoft Corporation||Data system with distributed tree indexes and method for maintaining the indexes|
|US5832514 *||Jun 26, 1996||Nov 3, 1998||Microsoft Corporation||System and method for discovery based data recovery in a store and forward replication process|
|US5842207 *||Jul 10, 1992||Nov 24, 1998||Hitachi, Ltd.||Method for storing records of a distributed database by plural processors to provide a host processor with sorted records belonging to one of a plurality of key sections|
|US5941955 *||Aug 9, 1995||Aug 24, 1999||British Telecommunications Public Limited Company||Recovery of distributed hierarchical data access routing system upon detected failure of communication between nodes|
|US5991764 *||Mar 12, 1997||Nov 23, 1999||International Business Machines Corporation||Data structure specifying differing fan-in tree and fan-out tree computation patterns supporting a generic reduction object for data parallelism|
|US6009425 *||Aug 21, 1996||Dec 28, 1999||International Business Machines Corporation||System and method for performing record deletions using index scans|
|US6070254 *||Oct 17, 1997||May 30, 2000||International Business Machines Corporation||Advanced method for checking the integrity of node-based file systems|
|US6105062 *||Feb 26, 1998||Aug 15, 2000||Novell, Inc.||Method and system for pruning and grafting trees in a directory service|
|US6138123 *||Jul 15, 1997||Oct 24, 2000||Rathbun; Kyle R.||Method for creating and using parallel data structures|
|US6285916||Apr 28, 1997||Sep 4, 2001||United Parcel Serivce Of America, Inc.||Multi-stage parcel tracking system|
|US6532487||Dec 4, 1995||Mar 11, 2003||International Business Machines Corporation||Method and system for managing semaphores in an object-oriented multi-tasking computer system|
|US6622259||Jul 14, 2000||Sep 16, 2003||International Business Machines Corporation||Non-disruptive migration of coordinator services in a distributed computer system|
|US6675173 *||Dec 30, 1999||Jan 6, 2004||Ori Software Development Ltd.||Database apparatus|
|US7028022||Jul 28, 2000||Apr 11, 2006||International Business Machines Corporation||Heuristic-based conditional data indexing|
|US7215360 *||Apr 5, 2002||May 8, 2007||Triveni Digital, Inc.||Error propagation tree technology|
|US7287033||Mar 5, 2003||Oct 23, 2007||Ori Software Development, Ltd.||Efficient traversals over hierarchical data and indexing semistructured data|
|US7315871||Jan 19, 2005||Jan 1, 2008||International Business Machines Inc. Corporation||Method, system and program product for interning invariant data objects in dynamic space constrained systems|
|US7401092||Jun 26, 2003||Jul 15, 2008||Standbysoft Llc||Method and apparatus for exchanging sub-hierarchical structures within a hierarchical file system|
|US7437595 *||Feb 7, 2005||Oct 14, 2008||International Business Machines Corporation||Row fault detection system|
|US7451342||Feb 7, 2005||Nov 11, 2008||International Business Machines Corporation||Bisectional fault detection system|
|US7506196 *||Feb 7, 2005||Mar 17, 2009||International Business Machines Corporation||System and method for detecting a faulty object in a system|
|US7506197 *||Feb 7, 2005||Mar 17, 2009||International Business Machines Corporation||Multi-directional fault detection system|
|US7509448||Jan 5, 2007||Mar 24, 2009||Isilon Systems, Inc.||Systems and methods for managing semantic locks|
|US7529963 *||Feb 7, 2005||May 5, 2009||International Business Machines Corporation||Cell boundary fault detection system|
|US7551572||Oct 21, 2005||Jun 23, 2009||Isilon Systems, Inc.||Systems and methods for providing variable protection|
|US7571345||Aug 22, 2008||Aug 4, 2009||International Business Machines Corporation||Bisectional fault detection system|
|US7590652||Aug 18, 2006||Sep 15, 2009||Isilon Systems, Inc.||Systems and methods of reverse lookup|
|US7593938||Dec 22, 2006||Sep 22, 2009||Isilon Systems, Inc.||Systems and methods of directory entry encodings|
|US7669075||Aug 25, 2008||Feb 23, 2010||International Business Machines Corporation||Row fault detection system|
|US7676691||Aug 18, 2006||Mar 9, 2010||Isilon Systems, Inc.||Systems and methods for providing nonlinear journaling|
|US7680836||Aug 18, 2006||Mar 16, 2010||Isilon Systems, Inc.||Systems and methods for a snapshot of data|
|US7680842||Aug 18, 2006||Mar 16, 2010||Isilon Systems, Inc.||Systems and methods for a snapshot of data|
|US7685126||Nov 9, 2001||Mar 23, 2010||Isilon Systems, Inc.||System and methods for providing a distributed file system utilizing metadata to track information about data stored throughout the system|
|US7743033||Jul 19, 2007||Jun 22, 2010||Isilon Systems, Inc.||Systems and methods for providing a distributed file system utilizing metadata to track information about data stored throughout the system|
|US7747895||Jul 1, 2008||Jun 29, 2010||International Business Machines Corporation||Multi-directional fault detection system|
|US7752402||Aug 18, 2006||Jul 6, 2010||Isilon Systems, Inc.||Systems and methods for allowing incremental journaling|
|US7756898||Mar 31, 2006||Jul 13, 2010||Isilon Systems, Inc.||Systems and methods for notifying listeners of events|
|US7779048||Apr 13, 2007||Aug 17, 2010||Isilon Systems, Inc.||Systems and methods of providing possible value ranges|
|US7788303||Oct 21, 2005||Aug 31, 2010||Isilon Systems, Inc.||Systems and methods for distributed system scanning|
|US7797283||Oct 21, 2005||Sep 14, 2010||Isilon Systems, Inc.||Systems and methods for maintaining distributed data|
|US7822932||Aug 18, 2006||Oct 26, 2010||Isilon Systems, Inc.||Systems and methods for providing nonlinear journaling|
|US7840834||Jul 1, 2008||Nov 23, 2010||International Business Machines Corporation||Multi-directional fault detection system|
|US7844617||Jun 4, 2010||Nov 30, 2010||Isilon Systems, Inc.||Systems and methods of directory entry encodings|
|US7848261||Feb 17, 2006||Dec 7, 2010||Isilon Systems, Inc.||Systems and methods for providing a quiescing protocol|
|US7853820||Oct 22, 2008||Dec 14, 2010||International Business Machines Corporation||System and method for detecting a faulty object in a system|
|US7870345||Mar 27, 2008||Jan 11, 2011||Isilon Systems, Inc.||Systems and methods for managing stalled storage devices|
|US7882068||Aug 21, 2007||Feb 1, 2011||Isilon Systems, Inc.||Systems and methods for adaptive copy on write|
|US7882071||Aug 18, 2006||Feb 1, 2011||Isilon Systems, Inc.||Systems and methods for a snapshot of data|
|US7899800||Aug 18, 2006||Mar 1, 2011||Isilon Systems, Inc.||Systems and methods for providing nonlinear journaling|
|US7900015||Apr 13, 2007||Mar 1, 2011||Isilon Systems, Inc.||Systems and methods of quota accounting|
|US7917474 *||Oct 21, 2005||Mar 29, 2011||Isilon Systems, Inc.||Systems and methods for accessing and updating distributed data|
|US7930584||Oct 9, 2008||Apr 19, 2011||International Business Machines Corporation||Cell boundary fault detection system|
|US7937421||Nov 14, 2003||May 3, 2011||Emc Corporation||Systems and methods for restriping files in a distributed file system|
|US7949636||Mar 27, 2008||May 24, 2011||Emc Corporation||Systems and methods for a read only mode for a portion of a storage system|
|US7949692||Aug 21, 2007||May 24, 2011||Emc Corporation||Systems and methods for portals into snapshot data|
|US7953704||Aug 18, 2006||May 31, 2011||Emc Corporation||Systems and methods for a snapshot of data|
|US7953709||Mar 27, 2008||May 31, 2011||Emc Corporation||Systems and methods for a read only mode for a portion of a storage system|
|US7962779||Jun 9, 2008||Jun 14, 2011||Emc Corporation||Systems and methods for a distributed file system with data recovery|
|US7966289||Aug 21, 2007||Jun 21, 2011||Emc Corporation||Systems and methods for reading objects in a file system|
|US7971021||Dec 16, 2010||Jun 28, 2011||Emc Corporation||Systems and methods for managing stalled storage devices|
|US7979735 *||Aug 15, 2008||Jul 12, 2011||International Business Machines Corporation||Data storage with snapshot-to-snapshot recovery|
|US7984137 *||May 10, 2006||Jul 19, 2011||Cisco Technology, Inc.||Methods and apparatus for maintaining a map of node relationships for a network|
|US7984324||Mar 27, 2008||Jul 19, 2011||Emc Corporation||Systems and methods for managing stalled storage devices|
|US8005865||May 27, 2010||Aug 23, 2011||Emc Corporation||Systems and methods for notifying listeners of events|
|US8010493||Mar 4, 2010||Aug 30, 2011||Emc Corporation||Systems and methods for a snapshot of data|
|US8015156||Mar 4, 2010||Sep 6, 2011||Emc Corporation||Systems and methods for a snapshot of data|
|US8015216||Jul 8, 2010||Sep 6, 2011||Emc Corporation||Systems and methods of providing possible value ranges|
|US8027984||Sep 4, 2009||Sep 27, 2011||Emc Corporation||Systems and methods of reverse lookup|
|US8028006||Jul 15, 2008||Sep 27, 2011||Standoffbysoft Llc||Method and apparatus for exchanging sub-hierarchical structures within a hierarchical file system|
|US8051425||Oct 28, 2005||Nov 1, 2011||Emc Corporation||Distributed system with asynchronous execution systems and methods|
|US8054765||Jun 15, 2009||Nov 8, 2011||Emc Corporation||Systems and methods for providing variable protection|
|US8055711||Oct 28, 2005||Nov 8, 2011||Emc Corporation||Non-blocking commit protocol systems and methods|
|US8060521||Nov 15, 2010||Nov 15, 2011||Emc Corporation||Systems and methods of directory entry encodings|
|US8065308||Oct 12, 2007||Nov 22, 2011||Ori Software Development Ltd.||Encoding semi-structured data for efficient search and browsing|
|US8073881 *||Aug 3, 2009||Dec 6, 2011||Sanbolic, Inc.||Methods and apparatus facilitating access to storage among multiple computers|
|US8082379||Mar 23, 2009||Dec 20, 2011||Emc Corporation||Systems and methods for managing semantic locks|
|US8112395||May 25, 2010||Feb 7, 2012||Emc Corporation||Systems and methods for providing a distributed file system utilizing metadata to track information about data stored throughout the system|
|US8112658||Aug 25, 2008||Feb 7, 2012||International Business Machines Corporation||Row fault detection system|
|US8117502||Aug 22, 2008||Feb 14, 2012||International Business Machines Corporation||Bisectional fault detection system|
|US8140623||Jun 8, 2006||Mar 20, 2012||Emc Corporation||Non-blocking commit protocol systems and methods|
|US8176013||Dec 10, 2010||May 8, 2012||Emc Corporation||Systems and methods for accessing and updating distributed data|
|US8181065||Mar 2, 2010||May 15, 2012||Emc Corporation||Systems and methods for providing nonlinear journaling|
|US8195905||Jan 13, 2011||Jun 5, 2012||Emc Corporation||Systems and methods of quota accounting|
|US8200632||Jan 14, 2011||Jun 12, 2012||Emc Corporation||Systems and methods for adaptive copy on write|
|US8214334||Jul 15, 2010||Jul 3, 2012||Emc Corporation||Systems and methods for distributed system scanning|
|US8214400||Aug 24, 2010||Jul 3, 2012||Emc Corporation||Systems and methods for maintaining distributed data|
|US8238350||Oct 28, 2005||Aug 7, 2012||Emc Corporation||Message batching with checkpoints systems and methods|
|US8286029||Dec 21, 2006||Oct 9, 2012||Emc Corporation||Systems and methods for managing unavailable storage devices|
|US8356013||Dec 17, 2010||Jan 15, 2013||Emc Corporation||Systems and methods for a snapshot of data|
|US8356150||Sep 30, 2010||Jan 15, 2013||Emc Corporation||Systems and methods for providing nonlinear journaling|
|US8364726 *||Nov 7, 2011||Jan 29, 2013||Sanbolic, Inc.||Methods and apparatus facilitating access to storage among multiple computers|
|US8380689||Feb 25, 2011||Feb 19, 2013||Emc Corporation||Systems and methods for providing nonlinear journaling|
|US8381022||Apr 18, 2011||Feb 19, 2013||International Business Machines Corporation||Data storage with snapshot-to-snapshot recovery|
|US8489597||Sep 1, 2004||Jul 16, 2013||Ori Software Development Ltd.||Encoding semi-structured data for efficient search and browsing|
|US8495411 *||Feb 7, 2005||Jul 23, 2013||International Business Machines Corporation||All row, planar fault detection system|
|US8539056||Aug 2, 2006||Sep 17, 2013||Emc Corporation||Systems and methods for configuring multiple network interfaces|
|US8539279||Jan 2, 2013||Sep 17, 2013||International Business Machines Corporation||Data storage with snapshot-to-snapshot recovery|
|US8625464||Nov 1, 2010||Jan 7, 2014||Emc Corporation||Systems and methods for providing a quiescing protocol|
|US8966080||Apr 13, 2007||Feb 24, 2015||Emc Corporation||Systems and methods of managing resource utilization on a threaded computer system|
|US9058334 *||Feb 11, 2010||Jun 16, 2015||Emc Corporation||Parallel file system processing|
|US9165157||Jan 14, 2013||Oct 20, 2015||Citrix Systems, Inc.||Methods and apparatus facilitating access to storage among multiple computers|
|US9665573||Dec 5, 2012||May 30, 2017||Nuix Pty Ltd||Parallelization of electronic discovery document indexing|
|US9742867 *||Mar 23, 2017||Aug 22, 2017||Sas Institute Inc.||Network data retrieval|
|US9785700||Aug 7, 2013||Oct 10, 2017||Nuix Pty Ltd||Systems and methods for load-balancing by secondary processors in parallelized indexing|
|US20020158969 *||Apr 5, 2002||Oct 31, 2002||Gupta Jimmy Rohit||Error propagation tree technology|
|US20050033733 *||Sep 1, 2004||Feb 10, 2005||Ori Software Development Ltd.||Encoding semi-structured data for efficient search and browsing|
|US20060176826 *||Feb 7, 2005||Aug 10, 2006||International Business Machines Corporation||All-to-all sequenced fault detection system|
|US20060179268 *||Feb 7, 2005||Aug 10, 2006||International Business Machines Corporation||Row fault detection system|
|US20060179271 *||Feb 7, 2005||Aug 10, 2006||International Business Machines Corporation||Bisectional fault detection system|
|US20060179340 *||Feb 7, 2005||Aug 10, 2006||International Business Machines Corporation||System and method for detecting a faulty object in a system|
|US20060218301 *||May 10, 2006||Sep 28, 2006||Cisco Technology, Inc.||Methods and apparatus for maintaining a map of node relationships for a network|
|US20070094310 *||Oct 21, 2005||Apr 26, 2007||Passey Aaron J||Systems and methods for accessing and updating distributed data|
|US20080021907 *||Jul 19, 2007||Jan 24, 2008||Patel Sujal M||Systems and methods for providing a distributed file system utilizing metadata to track information about data stored throughout the system|
|US20080046445 *||Aug 18, 2006||Feb 21, 2008||Passey Aaron J||Systems and methods of reverse lookup|
|US20080046475 *||Aug 18, 2006||Feb 21, 2008||Anderson Robert J||Systems and methods for a snapshot of data|
|US20080065596 *||Oct 12, 2007||Mar 13, 2008||Ori Software Development Ltd.||Encoding semi-structured data for efficient search and browsing|
|US20080071733 *||Oct 2, 2007||Mar 20, 2008||Ori Software Development Ltd.||Efficient traversals over hierarchical data and indexing semistructured data|
|US20080126365 *||Aug 18, 2006||May 29, 2008||Fachan Neal T||Systems and methods for providing nonlinear journaling|
|US20080243773 *||Jun 9, 2008||Oct 2, 2008||Isilon Systems, Inc.||Systems and methods for a distributed file system with data recovery|
|US20080270852 *||Jul 1, 2008||Oct 30, 2008||International Business Machines Corporation||Multi-directional fault detection system|
|US20080288820 *||Jul 1, 2008||Nov 20, 2008||International Business Machines Corporation||Multi-directional fault detection system|
|US20080313217 *||Jul 15, 2008||Dec 18, 2008||Standbysoft Llc||Method and apparatus for exchanging sub-hierarchical structures within a hierarchical file system|
|US20080313506 *||Aug 22, 2008||Dec 18, 2008||International Business Machines Corporation||Bisectional fault detection system|
|US20080320329 *||Aug 25, 2008||Dec 25, 2008||International Business Machines Corporation||Row fault detection system|
|US20080320330 *||Aug 25, 2008||Dec 25, 2008||International Business Machines Corporation||Row fault detection system|
|US20090044052 *||Oct 9, 2008||Feb 12, 2009||International Business Machines Corporation||Cell boundary fault detection system|
|US20090052334 *||Oct 22, 2008||Feb 26, 2009||International Business Machines Corporation||System and method for detecting a faulty object in a system|
|US20100042791 *||Aug 15, 2008||Feb 18, 2010||International Business Machines Corporation||Data storage with snapshot-to-snapshot recovery|
|US20100318835 *||Aug 22, 2008||Dec 16, 2010||International Business Machines Corporation||Bisectional fault detection system|
|US20110196841 *||Apr 18, 2011||Aug 11, 2011||International Business Machines Corporation||Data storage with snapshot-to-snapshot recovery|
|US20110196899 *||Feb 11, 2010||Aug 11, 2011||Isilon Systems, Inc.||Parallel file system processing|
|US20120011144 *||Dec 23, 2010||Jan 12, 2012||Frederik Transier||Aggregation in parallel computation environments with shared memory|
|US20170139782 *||Nov 16, 2015||May 18, 2017||Red Hat, Inc.||Recreating a computing environment using tags and snapshots|
|CN100422944C||Sep 8, 2004||Oct 1, 2008||张 健;孙汉翔;沈 文||Method of continuous recording and reserving data changing by data map|
|U.S. Classification||714/4.1, 707/E17.032, 707/E17.012, 714/E11.113|
|International Classification||G06F17/30, G06F12/00, G06F15/173, G06F11/14|
|Cooperative Classification||G06F11/1402, G06F17/30091|
|European Classification||G06F17/30F2, G06F11/14A|
|Apr 16, 1990||AS||Assignment|
Owner name: INTERNATIONAL BUSINESS MACHINES CORPORATION, A COR
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNORS:FREY, ALEXANDER H. JR.;MOSTELLER, RICHARD C.;REEL/FRAME:005282/0820
Effective date: 19900409
|Feb 15, 1994||CC||Certificate of correction|
|Sep 25, 1996||FPAY||Fee payment|
Year of fee payment: 4
|Feb 13, 2001||REMI||Maintenance fee reminder mailed|
|Jul 22, 2001||LAPS||Lapse for failure to pay maintenance fees|
|Sep 25, 2001||FP||Expired due to failure to pay maintenance fee|
Effective date: 20010720